Intracranial EEG-fMRI Analysis of Interictal Markers in Drug-Resistant Focal Epilepsy

Epilepsy is a common neurologic disease affecting 70 million people globally. Despite advances in treatment, approximately 20 million people remain resistant to medication for whom surgery may be considered. Epilepsy surgery requires the precise localization of the epileptogenic zone (EZ) and conventional markers such as interictal epileptiform discharges (IEDs) have been shown to provide incomplete spatial information to this end. Recently, high frequency oscillations (HFOs) in intracranial EEG (iEEG) have been proposed as a novel marker of the EZ. However, iEEG recordings are limited to electrode contact sites. To overcome this limitation, the Federico laboratory has combined iEEG with functional MRI which allows us to measure activation throughout the whole brain. This thesis contains a series of experiments using iEEG-fMRI in a cohort of patients with drug-resistant focal epilepsy with the goal of localizing the EZ using IEDs and HFOs and predicting surgical outcome. I first established methods that recovered the underlying physiological EEG that was heavily contaminated by magnetic gradient switching artifact and then detected HFOs algorithmically. I then validated the use of these methods showing that the rates of HFOs obtained from within the scanner were comparable to rates obtained at the bedside. The semi-automatic HFO detection method is a feasible tool with minimal user effort for the identification of HFOs in iEEG-fMRI data and is easily applicable in a clinical setting. Having validated the technique, I then used simultaneous iEEG-fMRI to produce HFO-related fMRI activation maps to examine their proximity to the seizure onset zone (SOZ). I evaluated 3 markers: IEDs, HFOs that co-occur with IEDs, and HFOs that occur independently of IEDs. I found HFOs that co-occur with IEDs produced discrete maps and co-localized with the SOZ better than HFOs that occurred independently of IEDs, which produced diffuse and non-localizing activations. Finally, I evaluated the relationship between resection of HFO-related fMRI activation and post-operative outcome against the use of conventional EEG marker, IEDs. I demonstrated that the resection of the peak voxel of the maximum cluster was necessary but not sufficient to achieve good post-operative outcomes for both IEDs and HFOs. The results of this thesis demonstrate the clinical value of HFOs and the utility of iEEG-fMRI as both provide useful information during pre-surgical evaluation